Why AI data centres are powering a boom in fuel-cell demand

Artificial intelligence data centers are driving demand for fuel cells as hyperscalers face grid constraints in securing power, according to industry projections. Rystad Energy anticipates investments in fuel cells for data centers to reach $30 billion by 2030, up from $2.8 billion in 2025. Fuel cells generate electricity via a chemical reaction between a fuel, like natural gas or hydrogen, and oxygen, bypassing combustion.

Bloom Energy and FuelCell Energy are at the forefront of this technology, with Bloom focusing on solid-oxide systems compatible with pure hydrogen and FuelCell on molten carbonate technology with a maximum 40% hydrogen blend. The rapid deployment of fuel cells is driven by the need for reliable power in data centers, which can take up to six years to secure from the grid.

Oracle has committed to purchasing up to 2.8 gigawatts of Bloom's fuel-cell systems, while FuelCell targets a significant portion of its future business in data centers. Despite promising growth, the fuel cell industry must scale up its manufacturing capacity to meet demand, with annual production expected to increase to 4 GW by 2030 from the current 1.8 GW.

Factors like site readiness, permitting, and supporting infrastructure can delay project timelines, but fuel cells offer a complementary power source to the grid, albeit in competition with other on-site generation technologies.